|Publication number||US7386767 B1|
|Application number||US 10/958,447|
|Publication date||Jun 10, 2008|
|Filing date||Oct 5, 2004|
|Priority date||Oct 5, 2004|
|Publication number||10958447, 958447, US 7386767 B1, US 7386767B1, US-B1-7386767, US7386767 B1, US7386767B1|
|Inventors||Ning Xue, Chong H Lee|
|Original Assignee||Altera Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (14), Referenced by (10), Classifications (14), Legal Events (3)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This invention relates to a bit error rate monitor for a serial communication interface. More particularly, this invention relates to a bit error rate monitor that is programmable to allow adjustment by a user to accommodate different interfaces.
Monitoring the bit error rate is important in serial communications applications. For example, the telecommunications industry closely monitors bit error rates. At various thresholds of bit error rate, a communications device may be instructed to slow down its transmission rate (and to return to its full rate when the error rate decreases), to request maintenance, to shut down or enter an idle mode, and/or to activate an alternate or back-up device.
Different communications devices have different error threshold requirements (i.e., different monitoring period and threshold levels). For example, the 10-Gigabit Ethernet 64b/66b padded protocol specification calls for monitoring two bits out of every 66 bits, and sets an error threshold of 10−4 (one bit error in ten thousand bits), while most telecommunications applications have a base error threshold of 10−3 (one bit error in one thousand bits) during a 125 μs period (after which the total bit counter and the error counter are reset). At 10 Gb/s, each bit period is 0.1 ns, so that 125 μs, which is 125,000 ns, translates to 1,250,000 bits. Thus the base bit error rate (one error per 1,000 bits) for telecommunications applications at 10 Gb/s is 1,250 errors in 1,250,000 bits, or 12.5 errors in 12,500 bits. Such applications also frequently report errors at additional thresholds of 10−6 (one error per 1,000,000 bits) and 10−9 (one error per 1,000,000,000 bits). It should be noted that, as is commonly the case in digital systems, these thresholds are actually multiples of 2−10 (one-per-1,024) rather than multiples of 10−3 (one-per-1,000), so that one-per-million threshold is actually 2−20 (or one-per-1,048,576) and the one-per-billion threshold is actually 2−30 (or one-per-1,073,741,824).
Bit error rate monitors exist for such applications and protocols. However, unlike in standard telecommunications applications, in data communications applications the user may define his or her own threshold value and/or base measurement period. Moreover, when users implement designs in programmable logic devices, those designs may deviate from standard clock rates—even when implementing known standards—in order to meet requirements of a particular implementation. In such a case, the standard bit error monitors may not be appropriate. For example, instead of using the 10 Gb/s standard described above, a user may implement an 11 Gb/s data rate. At 11 Gb/s, a 10−3 bit error threshold would translate to 13.75 bit errors per 13,750 bits instead of 12.5 errors per 12,500 bits, with similar adjustments for 10−6 and 10−9 thresholds. In such a case, the standard bit error monitors may not be appropriate.
The present invention provides a programmable bit error rate monitor. Although, as discussed above, one motivation for provision of a programmable error rate monitor is that users may implement “standard” interfaces in programmable logic devices (“PLDs”) in non-standard ways, so that different monitoring parameters may be necessary, preferably the programmable bit error rate monitor of the invention is provided separately from the programmable logic device on which the interface to be monitored is implemented. This is because the interface to be monitored normally would alter the data stream before it could be sent to the monitor. However, it may be possible to implement the programmable bit error rate monitor of the invention in the programmable logic core of the same PLD that includes the interface to be monitored. Alternatively, the programmable bit error rate monitor can be provided on a separate PLD. Either way, the programming to create the programmable bit error rate monitor may be provided as a preprogrammed “core” by the PLD supplier. In addition, the programmable bit error rate monitor can be provided as a discrete device with user-addressable memory into which the user could store the programmable parameters.
The programmable bit error rate monitor of the invention preferably includes a bit error counter that preferably monitors the data stream, and a monitoring period counter with a user-programmable upper bound. The bit error counter preferably outputs a signal to an error flag generator each time a bit error is detected. The monitoring period counter preferably outputs a reset signal when the upper bound, which defines the monitoring period, is reached.
The error flag generator preferably includes a programmable threshold register and a register for accumulating the bit error counts. Upon receipt of the reset signal from the monitoring period counter, the bit error generator preferably compares the bit error count to the programmed threshold stored in the threshold register, and generates at least one error flag when the bit error count exceeds the threshold. In one preferred embodiment, the error flag generator preferably generates three error flags, representing the three thresholds (10−3, 10−6 and 10−9) discussed above. In such an embodiment, preferably a multiplexer is provided to allow a user to programmably select which error flag is output by the bit error rate monitor.
It would be expected that a user would program the monitoring period counter and the error threshold of the error flag generator so that an error flag is generated when an error rate that the user's application cannot tolerate is reached. Moreover, the user would be expected to select the sensitivity (e.g., 10−3, 10−6 and 10−9) required by the user's application. Normally, one would expect the user to monitor the highest sensitivity, changing to a lower sensitivity only on occurrence of an error flag, to see just how bad the error condition is.
Thus, a user might monitor only the 10−9 flag under normal conditions, and as long as that flag is not asserted, everything is normal. If the 10−9 flag is asserted, however, it means there is more than one error per billion bits, but that could include a situation where there is more than one error per million bits, which might require throughput reduction or tighter flow control, and even more than one error per thousand bits, which might require a system shutdown. So in a case where the 10−9 is asserted, the user might then switch to the 10−6 flag. If the 10−6 flag is not asserted, then the user knows that the noise level is still acceptable. If the 10−6 also is asserted, then the user might switch to the 10−3 flag. If the 10−3 flag is asserted, drastic action such as shutdown may be required, but if the 10−3 is not asserted, then the user knows that the condition is in the 10−6 range, and can take less drastic action.
The above and other advantages of the invention will be apparent upon consideration of the following detailed description, taken in conjunction with the accompanying drawings, in which like reference characters refer to like parts throughout, and in which:
The invention will now be described with reference to
As shown in
A clock signal 16 associated with data signal 14, which may be provided separately from data signal 14 or may be recovered from data signal 14 by clock recovery circuitry (not shown) as is well known, is input to programmable monitoring period counter 12. Programmable monitoring period counter 12 preferably includes a user-programmable memory 120 into which a user, via input 121, can enter an upper bound representing the duration of the monitoring period. Programmable monitoring period counter 12 preferably increments once per cycle of clock 11. When the count of programmable monitoring period counter 12 reaches the user-programmed upper bound, it preferably asserts a signal at 122, which is output to both bit error counter 11 and error flag generator 13, which use signal 122 as described below.
Error flag generator 13 preferably includes comparator(s) 130 and a user-programmable memory 131 into which a user, via input 132, can enter an error threshold value. The value in user-programmable memory 131 is one input to the comparator, while error count signal 15 is the other input to the comparator. Signal 122 is used as an enable signal for the comparator.
When the monitoring period duration is reached, monitoring period counter 12 briefly asserts signal 122, which enables the comparator(s) in error flag generator 13. If, while a comparator 130 is enabled, the error count 15 exceeds the threshold value in memory 131, then error flag generator 13 asserts an error flag 133. Signal 122 also functions as a reset signal for bit error counter 11, so that the error count returns to zero for the start of a new monitoring period.
Error flag generator 13 preferably provides not only basic error flag 133, but preferably also 1,000-times-error flag 134 and 1,000,000-times-error flag 135. Thus, if the basic error flag 133 represents one error per billion bits, then 1,000-times-error flag 134 represents one error in one million bits and 1,000,000-times-error flag 135 represents one error in one thousand bits. As stated above, these flags 133-135 are actually multiples of 1,024 rather than multiples of 1,000, so that 1,000-times-error flag 134 is actually 1,024 times less sensitive than error flag 133, while 1,000,000-times-error flag 135 is actual 1,0242 times (or 1,048,576 times) less sensitive than error flag 133.
Flag selector 136, which preferably is a multiplexer as shown, is programmable by user input 137 to select one of the three flags 133, 134, 135 as the output 138 of error flag generator 13.
It is possible that error flag generator 13 makes three separate comparisons to generate flags 133-135. In one embodiment of such a case, the user might program three separate thresholds in memory or memories 131, and a separate comparison would be made between error count signal 15 and each threshold. However, preferably 1,000-times-error flag 134 and 1,000,000-times-error flag 135 are extrapolated from base error flag 133. One way that this can be done is shown in
As seen in
For the 1,000-times comparison shown in
For the 1,000,000-times comparison shown in
It follows from the foregoing that in most cases, a 10-bit error threshold is sufficient, considering that for flags 134 and 135, the ten or twenty most significant bits of the number in threshold memory 131 are ignored. Indeed, in a preferred embodiment, if a user programs a threshold value into threshold memory 131 that has any ones in the twenty most significant bits then flag 135 is not available, and if any of those ones are in the ten most significant bits then flag 134 also is not available.
The discussion so far has assumed a monitoring period on the order of one second. However, greater error resolution can be obtained by lengthening the monitoring period. A factor of ten increase in the duration of monitoring period results in substantially a factor of ten increase in resolution. Therefore, by lengthening the monitoring period sufficiently, the bit error rate can be measured with a resolution of 10−18 to 10−15, which is in the range of error rates for many telecommunications applications.
The present invention provides users with the flexibility to adjust a bit error rate monitor to accommodate any deviations in their designs from known standards as described above by allowing easy adjustment of the monitoring period and the error threshold, as well as the easy selection of error flags or different sensitivities. Thus, the hypothetical user described above who implements an 11 Gb/s interface can easily adapt bit error rate monitor 10 to accommodate that interface.
As stated above, bit error rate monitor 10 according to the present invention may be implemented in a dedicated circuit having programmable memories for monitoring period upper bound memory 120 and threshold memory 131. Alternatively, bit error rate monitor 10 could be implemented in a programmable logic device. Either way, as seen in
PLD 50 with which, or in which, bit error rate monitor 10 according to the present invention may be used, preferably is programmably configurable to handle any of a plurality of high-speed communication protocols. A PLD 50 incorporating a bit error rate monitor according to the present invention may be used in many kinds of electronic devices. One possible use is in a data processing system 900 shown in
System 900 can be used in a wide variety of applications, such as computer networking, data networking, instrumentation, video processing, digital signal processing, or any other application where the advantage of using programmable or reprogrammable logic is desirable. PLD 50 can be used to perform a variety of different logic functions. For example, PLD 50 can be configured as a processor or controller that works in cooperation with processor 901. PLD 50 may also be used as an arbiter for arbitrating access to a shared resources in system 900. In yet another example, PLD 50 can be configured as an interface between processor 901 and one of the other components in system 900. It should be noted that system 900 is only exemplary, and that the true scope and spirit of the invention should be indicated by the following claims.
Various technologies can be used to implement PLDs 50 as described above and incorporating this invention.
It will be understood that the foregoing is only illustrative of the principles of the invention, and that various modifications can be made by those skilled in the art without departing from the scope and spirit of the invention, and the present invention is limited only by the claims that follow.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US4447902 *||Mar 19, 1982||May 8, 1984||Sony Corporation||Error-correction format for digital television signals|
|US4451916 *||May 12, 1980||May 29, 1984||Harris Corporation||Repeatered, multi-channel fiber optic communication network having fault isolation system|
|US5325397 *||Dec 6, 1990||Jun 28, 1994||The Commonwealth Of Australia||Error rate monitor|
|US5606563 *||May 18, 1995||Feb 25, 1997||Pmc-Sierra, Inc.||Programmable jump window for sonet compliant bit error monitoring|
|US5623497 *||Feb 9, 1995||Apr 22, 1997||Advantest Corporation||Bit error rate measurement apparatus|
|US5764651 *||May 17, 1996||Jun 9, 1998||Integrated Device Technology, Inc.||Bit error rate detection system|
|US6973600 *||Feb 1, 2002||Dec 6, 2005||Adc Dsl Systems, Inc.||Bit error rate tester|
|US6988227 *||Jun 25, 2001||Jan 17, 2006||Silicon Laboratories Inc.||Method and apparatus for bit error rate detection|
|US7194666 *||Dec 1, 2003||Mar 20, 2007||Altera Corporation||Bit error rate tester implemented in a programmable logic device|
|US7231558 *||Apr 24, 2003||Jun 12, 2007||Finisar Corporation||System and method for network error rate testing|
|US7234086 *||Jan 13, 2004||Jun 19, 2007||Pmc Sierra, Inc.||Overlapping jumping window for SONET/SDH bit error rate monitoring|
|US20040083077 *||Oct 9, 2003||Apr 29, 2004||Broadcom Corporation||Integrated packet bit error rate tester for 10G SERDES|
|US20060020412 *||Jul 23, 2004||Jan 26, 2006||Bruensteiner Matthew M||Analog waveform information from binary sampled measurements|
|US20060069971 *||Sep 14, 2005||Mar 30, 2006||Waschura Thomas E||Method and apparatus for deferred decision signal quality analysis|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US7617424 *||Dec 30, 2005||Nov 10, 2009||Intel Corporation||Error monitoring for serial links|
|US7676701 *||Nov 30, 2006||Mar 9, 2010||Fuji Xerox Co., Ltd.||Computer readable medium storing an error recovery program, error recovery method, error recovery apparatus, and computer system|
|US8195996||May 5, 2008||Jun 5, 2012||Intel Corporation||Apparatuses, systems and methods for detecting errors in data streams on point-to-point serial links|
|US8627156 *||Oct 26, 2010||Jan 7, 2014||Agilent Technologies, Inc.||Method and system of testing bit error rate using signal with mixture of scrambled and unscrambled bits|
|US9032261||Apr 24, 2013||May 12, 2015||Skymedi Corporation||System and method of enhancing data reliability|
|US20070157054 *||Dec 30, 2005||Jul 5, 2007||Timothy Frodsham||Error monitoring for serial links|
|US20080010483 *||Nov 30, 2006||Jan 10, 2008||Fuji Xerox Co., Ltd.||Computer readable medium storing an error recovery program, error recovery method, error recovery apparatus, and computer system|
|US20080137719 *||Jun 26, 2007||Jun 12, 2008||Fujitsu Limited||Serial transmission system, a method for automatically correcting signal quality of serial transmission system, and port|
|US20080209306 *||May 5, 2008||Aug 28, 2008||Timothy Frodsham||Error monitoring for serial links|
|US20100031096 *||Jul 31, 2008||Feb 4, 2010||Ercole Rosario Di Iorio||Internal fail bit or byte counter|
|U.S. Classification||714/704, 714/715, 714/716, 714/725, 714/48, 714/707, 714/705, 714/706, 714/712|
|Cooperative Classification||G06F11/0772, G06F11/076|
|European Classification||G06F11/07P2A2, G06F11/07P4B|
|Oct 5, 2004||AS||Assignment|
Owner name: ALTERA CORPORATION, CALIFORNIA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:XUE, NING;LEE, CHONG H.;REEL/FRAME:015877/0194;SIGNING DATES FROM 20040921 TO 20040928
|Sep 23, 2011||FPAY||Fee payment|
Year of fee payment: 4
|Nov 24, 2015||FPAY||Fee payment|
Year of fee payment: 8